| Title: | ITALICS |
|---|---|
| Description: | A Method to normalize of Affymetrix GeneChip Human Mapping 100K and 500K set |
| Authors: | Guillem Rigaill, Philippe Hupe |
| Maintainer: | Guillem Rigaill <[email protected]> |
| License: | GPL-2 |
| Version: | 2.67.0 |
| Built: | 2024-11-29 07:14:07 UTC |
| Source: | https://github.com/bioc/ITALICS |
This function merge information obtain from the getQuartet function and a given table
addInfo(quartet, dat)addInfo(quartet, dat)
quartet |
list obtain through the getQuartet Function |
dat |
a data.frame with additionnal information it must contain a fsetid and fid column |
a data.frame similar to the quartetInfo item of quartet plus additionnal column
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
Glad Analysis ot the genomic profile
analyseCGH(data, amplicon, deletion, deltaN, forceGL, param, nbsigma, ...)analyseCGH(data, amplicon, deletion, deltaN, forceGL, param, nbsigma, ...)
data |
A data frame containing SNP's intensity, chromosome and position on the genome. data must have a Chr, X and LogRatio columns |
amplicon |
see the amplicon parameter in the daglad function |
deletion |
see the deletion parameter in the daglad function |
deltaN |
see the deltaN parameter in the daglad function |
forceGL |
see the forceGL parameter in the daglad function |
param |
see the param parameter in the daglad function |
nbsigma |
see the nbsigma parameter in the daglad function |
... |
Other daglad parameters. |
An object of class profileCGH
People interested in tools dealing with array CGH analysis and DNA copy number analysis can
visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function removes the LogRatio column of the snpInfo data.frame. Then compute the copy number of each SNP having its quartet intensities. And return the snpInfo data.frame with the newly computed LogRatio.
fromQuartetToSnp(quartetInfo, snpInfo, cIntensity="quartetLogRatio", nLog=1)fromQuartetToSnp(quartetInfo, snpInfo, cIntensity="quartetLogRatio", nLog=1)
quartetInfo |
A table containing the quartet intensities and other quartet information. It must have a column called : fsetid. |
snpInfo |
A table containing snp information. |
cIntensity |
A vector containing the names of the quartet information to be aggregate. For example quartetLogRatio. |
nLog |
The position of the field which will be named LogRatio in the snpInfo data.frame. For example if cIntensity = c("a", "b") and you want b to be considered as the LogRatio you should set nLog=2 |
return the data.frame snpInfo with additionnal columns.
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function put the smoothing value of each SNP in front of its corresponding quartet in the quartetInfo data.frame.
fromSnpToQuartet(quartetInfo, profilSNP)fromSnpToQuartet(quartetInfo, profilSNP)
quartetInfo |
a data frame containing all the quartet values plus there GC content, fragment length and GC content and Quartet effect |
profilSNP |
a data frame, corresponding to the profileValues argument of a profilCGH object (see GLAD) |
return the data.frame quartetInfo with an additionnal column: "Smoothing" corresponding to the estimated smoothing value.
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function eliminate badly predicted probes using a regression table and an estimated model given by the function getModel or getBestBICModelLight. Then it computes the corrected intensity.
getConfDat(confidence, quartetInfo, model)getConfDat(confidence, quartetInfo, model)
confidence |
The confidence interval : 0.95 |
quartetInfo |
A Regression table containing the variables in the model |
model |
The class lm object given by the function |
A data frame with the corrected intensity. Only goodly predicted probes are taken into account. SNP's with more than 8 badly predicted probes get a NA.
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function computes the corrected intensity.
getCorrection(effet, model, regTab)getCorrection(effet, model, regTab)
effet |
The name of the biological effect |
model |
The class lm object given by the |
regTab |
The regression table used to estimate the linear model, and containing the variables in the model |
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function retrieves the estimated biological effect
getEffet(effet, model, regTab)getEffet(effet, model, regTab)
effet |
The name of the biological effect |
model |
The class lm object given by the |
regTab |
The regression table used to estimate the linear model, and containing the variables in the model |
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
Computes the linear regression model and and return an object of class lm.
getModel(formule, response, regTab)getModel(formule, response, regTab)
formule |
A symbolic description of the term of the model. It is a string |
response |
The parameter you want to explain (the response) : the SNP "LogRatio". Y is a string |
regTab |
A Regression table containing the variables in the model |
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function retrieve information of each quartet. This function use the pd.mapping50k.xba240, pd.mapping50k.hind240, pd.mapping250k.sty and pd.mapping250k.nsp package.
getQuartet(pkgname, snpInfo)getQuartet(pkgname, snpInfo)
pkgname |
the chip type pd.mapping50k.xba240, pd.mapping50k.hind240, pd.mapping250k.sty or pd.mapping250k.nsp |
snpInfo |
a data frame containing SNPs position along the genome |
return a list with two fields. fid : containing the position of each quartet on the CEL file. quartetInfo : a data fame containing the columns : fsetid, fid, FL (fragment length) and GC (content of the quartet)
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function retrieves the residual values
getResidu(model)getResidu(model)
model |
The class lm object given by the |
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
This function retrieve the chromosome and position in bp of each SNP of a given Affymetrix SNP array. This function use the pd.mapping50k.xba240, pd.mapping50k.hind240, pd.mapping250k.sty and pd.mapping250k.nsp package.
getSnpInfo(pkgname)getSnpInfo(pkgname)
pkgname |
the chip type pd.mapping50k.xba240, pd.mapping50k.hind240, pd.mapping250k.sty or pd.mapping250k.nsp |
Return a data.frame with five columns : fsetid, dbsnp_rs_id, Chr, X and fragment_length corresponding to the fsetid, the rs_id, the chromosome, the position on the chromosome and the PCR amplified fragment length respectively.
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
Normalize and analyse Affymetrix SNP array 100K and 500K set (see the vignette)
ITALICS(quartetInfo, snpInfo, confidence=0.95, iteration=2, formule="Smoothing+QuartetEffect+FL+I(FL^2)+I(FL^3)+GC+I(GC^2)+I(GC^3)", prc=0.3, amplicon=2.1, deletion=-3.5, deltaN=0.15, forceGL=c(-0.2,0.2), param=c(d=2), nbsigma=1, ... )ITALICS(quartetInfo, snpInfo, confidence=0.95, iteration=2, formule="Smoothing+QuartetEffect+FL+I(FL^2)+I(FL^3)+GC+I(GC^2)+I(GC^3)", prc=0.3, amplicon=2.1, deletion=-3.5, deltaN=0.15, forceGL=c(-0.2,0.2), param=c(d=2), nbsigma=1, ... )
quartetInfo |
a data frame containing all the raw quartet intensities plus there GC content, fragment length, and Quartet effect |
snpInfo |
a data frame containing SNPs position along the genome and raw copy number |
confidence |
The confidence interval. After the last bias estimation step, quartets outside this confidence interval are flagged. The lower confidence is, the more quartets will be flagged. See also the parameter prc. |
iteration |
The number of iteration you d'like to do |
formule |
A symbolic description of the term of the model. The default value of formule means that we want correct the observed quartetLogRatio using the estimated copy number (Smoothing), the Quartet Effect, the quartet Fragment Length (FL) and the quartet GC content. |
prc |
prc is a frequence (between 0 and 1). After the final iteration of ITALICS, badly predicted probes are flagged (see also the parameter confidence). Only SNPs having more than prc of their probes non-flagged are kept for the final GLAD analysis. The higher prc is, the more SNPs are removed before the final GLAD analysis. |
amplicon |
see the amplicon parameter in the daglad function |
deletion |
see the deletion parameter in the daglad function |
deltaN |
see the deltaN parameter in the daglad function |
forceGL |
see the forceGL parameter in the daglad function |
param |
see the param parameter in the daglad function |
nbsigma |
see the nbsigma parameter in the daglad function |
... |
Other daglad parameters. |
The function ITALICS implements the methodology which
is described in the article : ITALICS: an algorithm for normalization and DNA copy number calling for Affymetrix SNP arrays (Rigaill et al., Bioinformatics Advance Access published on February 5, 2008).
The principle of the ITALICS algorithm: ITALICS, is a normalization method that estimates both the biological and the non-relevant effects in an alternate and iterative way to accurately remove the non-relevant effects.
ITALICS deals with known systematic sources of variation such as the GC-content of the quartets, the PCR amplified fragment length and the GC-content of the PCR amplified fragment . It also takes into account the quartet effect which corresponds to the fact that some quartets systematically have a small intensity while others tend to have a high intensity. ITALICS is also able to correct spatial artifacts which sometimes arise on Affymetrix SNP arrays 100K and 500K set.
Return an object of class profileCGH
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
## Not run: ## step to get the path of the HF0844_Hind.CEL file ITALICSDataPATH <- attr(as.environment(match("package:ITALICSData",search())),"path") filename <- paste(ITALICSDataPATH,"/extdata/HF0844_Hind.CEL", sep="") quartetEffectFile <- paste(ITALICSDataPATH,"/data/Hind.QuartetEffect.csv", sep="") ## load quartet effect quartetEffect <- read.table(quartetEffectFile, sep=";", header=TRUE) ## load annotation using the pd.mapping50k.xba24 or pd.mapping50k.hind240 or pd.mapping250k.sty or pd.mapping250k.nsp package headdetails <- readCelHeader(filename[1]) pkgname <- cleanPlatformName(headdetails[["chiptype"]]) snpInfo <- getSnpInfo(pkgname) quartet <- getQuartet(pkgname, snpInfo) ## read cel files and format data tmpExprs <- readCelIntensities(filename, indices=quartet$fid) quartet$quartetInfo$quartetLogRatio <- readQuartetCopyNb(tmpExprs) quartet$quartetInfo <- addInfo(quartet, quartetEffect) snpInfo <- fromQuartetToSnp(cIntensity="quartetLogRatio", quartetInfo=quartet$quartetInfo, snpInfo=snpInfo) ## ITALICS normalization profilSNPHind <- ITALICS(quartet$quartetInfo, snpInfo, formule="Smoothing+QuartetEffect+FL+I(FL^2)+I(FL^3)+GC+I(GC^2)+I(GC^3)") ## plot the profile data(cytoband) plotProfile(profilSNPHind, Smoothing="Smoothing", Bkp=TRUE, cytoband = cytoband) ## End(Not run)## Not run: ## step to get the path of the HF0844_Hind.CEL file ITALICSDataPATH <- attr(as.environment(match("package:ITALICSData",search())),"path") filename <- paste(ITALICSDataPATH,"/extdata/HF0844_Hind.CEL", sep="") quartetEffectFile <- paste(ITALICSDataPATH,"/data/Hind.QuartetEffect.csv", sep="") ## load quartet effect quartetEffect <- read.table(quartetEffectFile, sep=";", header=TRUE) ## load annotation using the pd.mapping50k.xba24 or pd.mapping50k.hind240 or pd.mapping250k.sty or pd.mapping250k.nsp package headdetails <- readCelHeader(filename[1]) pkgname <- cleanPlatformName(headdetails[["chiptype"]]) snpInfo <- getSnpInfo(pkgname) quartet <- getQuartet(pkgname, snpInfo) ## read cel files and format data tmpExprs <- readCelIntensities(filename, indices=quartet$fid) quartet$quartetInfo$quartetLogRatio <- readQuartetCopyNb(tmpExprs) quartet$quartetInfo <- addInfo(quartet, quartetEffect) snpInfo <- fromQuartetToSnp(cIntensity="quartetLogRatio", quartetInfo=quartet$quartetInfo, snpInfo=snpInfo) ## ITALICS normalization profilSNPHind <- ITALICS(quartet$quartetInfo, snpInfo, formule="Smoothing+QuartetEffect+FL+I(FL^2)+I(FL^3)+GC+I(GC^2)+I(GC^3)") ## plot the profile data(cytoband) plotProfile(profilSNPHind, Smoothing="Smoothing", Bkp=TRUE, cytoband = cytoband) ## End(Not run)
This function read the cel files and return the raw-value of each quartet = mean of allele A and B
readQuartetCopyNb(tmpExprs)readQuartetCopyNb(tmpExprs)
tmpExprs |
A vector of the perfect match intensity of allele A and B of the quartets. This vector should be sorted in a specific order. See the example given in the help of the ITALICS function. |
return a vector with the raw-value of each quartet
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].
Estimation of the quartet effect based on several normal sample chips
trainITALICS (dir, amplicon=2.1, deletion=-3.5, deltaN=0.15, forceGL=c(-0.2,0.2), param=c(d=2), nbsigma=1, ...)trainITALICS (dir, amplicon=2.1, deletion=-3.5, deltaN=0.15, forceGL=c(-0.2,0.2), param=c(d=2), nbsigma=1, ...)
dir |
The directory containing the normal sample chips. All theses chips should be of the same type hind, xba, nsp or sty. Only .CEL files be considered |
amplicon |
see the amplicon parameter in the daglad function |
deletion |
see the deletion parameter in the daglad function |
deltaN |
see the deltaN parameter in the daglad function |
forceGL |
see the forceGL parameter in the daglad function |
param |
see the param parameter in the daglad function |
nbsigma |
see the nbsigma parameter in the daglad function |
... |
Other daglad parameters. |
The ITALICS function take into account a quartet effect which is computed on a reference data set of normal women samples. The ITALICSData provide quartetEffect for the Xba, Hind, Sty and Nsp chip computed on our own reference data set.
We recommand that you use your own reference data set to compute the quartet Effect by using the trainITALICS function. ITALICS reference data should contain only woman normal samples. Furthermore we recommand that you check that none of these chip have obvious spatial artifact. To so read the cel files using the read.affybatch (form the affy package). Then use the image function on the obtain affybatch object.
a data.frame with two column fsetid and quartetEffect
People interested in tools dealing with array CGH analysis and DNA copy number analysis can visit our web-page http://bioinfo.curie.fr.
Guillem Rigaill, [email protected].
Institut Curie, [email protected].